1. Rejected because you cannot create objects in global scope which is the same as init. line 22 (services/log.js). 2. You are filling an array with logs, that is unnecessary for most users. 3. GNOME 41 has `console.debug()` that you can use instead of log(). I think it's better to leave log() as it was before and when you added 41 support, switch to `console.log`.

You are giving me contradicting guidelines. In the previous review, you said not to use too many logs. So a deleted all log() calls. Now, i don't use any logs, and handle the logs myself. Also this way the user can copy the error message and send it to me (without opening any terminal). So right now your guideline basically suggest i shouldn't do any logging. However, I don't know how to fix bug reports without logs, and i also can't expect users to go into my plugins metadata.js and flip a debug flag AND open a terminal and copy the logs to me. So.... what is exactly is the right solution here? (Also, i don't see why gnome 41 matters here. I don't support it and most users still use <40 gnome versions.)

1. Yes, you shouldn't log like version 5 especially when you are logging the spawn command outputs. 2. By saying this: "I think it's better to leave log() as it was before", I mean use log() instead of creating array to store everything. As I said before, `log()` should only get triggered on debug mode.

1. I need the outputs of spawn command to debug. 2. "log()` should only get triggered on debug mode" how can an average user do that?

Hi, we were discussing this issue in the Extensions matrix channel (https://matrix.to/#/#extensions:gnome.org) The issue with printing excessively to the log is that it makes debugging problems difficult for other applications on the system, and typically the log should only be used for warning or errors. If you were only logging when the subprocess returned an error that would not be an issue, but you seem to unconditionally log the output. In your current approach, your `Log` object is a persistent singleton - it will continuously increase its memory usage without ever freeing any of it. I'd like you to consider that the output of the `ddcutil` command should be relatively static, or at least it should only be necessary to log that output when a user encounters a problem. In that case, you could execute and log the output of that command as a part of your "Show logging" dialog. Does that make sense?

1. "If you were only logging when the subprocess returned an error that would not be an issue" - okay. 2. In your current approach, your `Log` object is a persistent singleton - it will continuously increase its memory usage without ever freeing any of it. -true, fixing it.